...But seriously, fuck cloudflare

Heya!
So, about two years ago, i had to move off of expansehost, because those guys suck really badly at hosting things.
At the time, i was using wordpress for this site, because expansehost had an offer for wordpress hosting.
But… Well, wordpress kind of sucked. I was using it because it was there, not because it was good.

So, after months of downtime because expansehost couldnt get its shit together, i decided to switch both blogging software and server host.

And as it happened, there was this really cool project just getting started at the time:
Deno Deploy.

I saw the original presentation by Ryan Dahl, and was immediately sold:
Serverside Javascript with an actually useable standard library?
And building on regular web api’s, and not some arcane, node-specific bullshit?

I mean, sign me up, right? That sounded incredible!
Coincidentally, I’d also discovered astro around that time, which had a really, really nice blogging starter template, with all the bells and whistles you could want. And it worked with deno! And it was extensible too!

And, even better than that, deno was actually running a hosting platform with a very generous free tier!

So, I mean, the stars aligned, right?
I made an astro starter project, put it on deno deploy, everything was great.

…Until today.
Because, apparently, Cloudflare bought the company behind Deno.

At first, you’d think that’d be a good thing.
Cloudflare has lots of money, right? That means Deno’s future is secured!

Well… No.

Cloudflare didn’t buy Deno Land for deno. They bought them for their other project, celld, which is basically a way to self-host cloudflare workers.

Deno, the software, is dead. The runtime will be discontinued in a year, deno deploy will no longer be around in six months.

Sure, the deno runtime is open-source.
If people care enough, it’ll continue being maintained.
Still, that’s quite the uncertain future for something your project fundamentally depends on. The fact that they would essentially kill off an entire runtime just for celld seems so completely insane and irresponsible to me. Deno does things that no other JS runtime can really do, especially when it comes to their permission system and inline imports. I’m not affected much, but so many projects are going to be royally fucked by this decision.

Anyways, what’d i do? Is this blog gonna be dead in 6 months?

Well, no, obviously not. And I’m also not gonna pay cloudflare to migrate to CF pages.

Fuck them for killing deno, and for controlling like half the webs’ traffic.

Luckily, migrating all this garbage took me about 30 minutes. Because I run a static site, and I wasn’t using any deno-specific features.

And even if i was, i could just keep using an outdated runtime. None of my backend code ever touches the web.

I pulled down the repo on a personal kvm, changed around some nginx config files, edited some AA DNS entries, and that was about all i had to do.

For the sake of convenience i also set up auto deploys from github via a simple shell script and a systemd service.

So, The moral here is to not use any fancy managed hosting services, use a static site for your blog, and avoid vendor lock-in. Also, uh… hope cloudflare doesn’t buy whatever tech you’re using just to kill it off.

That’s it, cya!